The Scoop
The White House imposed export controls on Anthropic’s powerful Mythos AI model partly over suspicions that a China-linked group had accessed it, a person familiar with the matter said.
The Trump administration on Friday directed Anthropic to limit access to Mythos, and its consumer version, known as Fable 5, to US citizens. That meant that even Anthropic employees who are foreign nationals would have been prohibited from using the models. Anthropic chose to remove the models from the market completely.
Anthropic tightly controlled the release of Mythos, which was launched in April, limiting access to a select group of companies that could use it to plug security holes. The plan was to ensure its capabilities wouldn’t make it into the hands of hackers. Anthropic has said Mythos represents a danger to the public because of its ability to find bugs in computer code, which could be exploited by malicious actors.
It’s unclear how the White House learned of the issue, which organization accessed the model, and how it gained access to Mythos.
But if the Chinese government had access to Mythos, it could pose national security risks to the US. China could also attempt to reverse-engineer and copy the model in a process known as distillation.
On Saturday, an adviser to President Trump, David Sacks, in a post on X shed new light on the events leading up to the White House decision to implement the export controls.
Sacks said that the government received a warning that Fable 5 could be jailbroken. When the Trump Administration notified Anthropic of the jailbreak, Anthropic co-founder and CEO Dario Amodei said the jailbreak was not a serious risk and refused to fix it, Sacks alleged.
Anthropic released Fable 5 to the public earlier this week — a version of Mythos with guardrails that prevent its use for cybersecurity purposes. A jailbreak is a way to remove the guardrails of an AI model.
A person close to the White House said Amazon informed the government about the jailbreak, and that its CEO, Andy Jassy, had been in contact with members of the administration about it.
Sacks said the administration reluctantly issued the export controls, which required Anthropic to only allow US citizens access to Mythos and Fable. Those requirements caused Anthropic to revoke access to the models for everyone.
An Anthropic spokesperson said the White House didn’t raise Chinese access to Mythos in its conversations around the Fable jailbreak and export controls. The AI company prohibits access to its products from within China.
An Amazon spokesperson said “it’s not uncommon for governments to seek our counsel on potential security risks. When they occur, we don’t share the details of these discussions.”
The White House and Sacks did not immediately respond to requests for comment. The Information first reported that Jassy had contacted the White House.
Know More
“In the past, Anthropic has always said that safety must be top priority and taken super seriously. In this case, Anthropic prioritized the continued offering of the consumer model over safety,” Sacks wrote in his tweet.
It isn’t the first time Anthropic has clashed with the White House. It lobbied against the administration’s efforts to preempt state AI regulation, and it is suing the Pentagon after reaching an impasse over the use of its models for autonomous weapons.
Sacks says those previous conflicts did not factor into the administration’s decision on Mythos. “Those trying to misdirect and tie this action to the prior DoW/Anthropic issues are wrong. The Admin values Anthropic’s technical capabilities and feels that this issue, while serious, should be easily resolved. The ball is in Anthropic’s court,” he said in the X post.